Brněnské Pyvo – Asynchronní

In our company we’ve got lots of Python services developed using asyncio, which helps us to keep them scalable and maintainable. But with all its benefits the asynchronous approach brings some new challenges. I would like to talk about these challenges and common pitfalls, explain some points requiring attention, such as concurrency control, timeouts, task results, and what metrics can be used to monitor health of async services.

Max is a Senior Software Engineer from the FinTech Tribe at Kiwi.com. See his GitHub: https://github.com/maxpolezhaev

I'll talk about my experience with running system processes (commands, scripts) from Python as a packager and automation engineer. Most of the tools I've created, including apkg the cross-distro packaging automation tool, use various system tools through CLI so a robust way to invoke commands is necessary.

Especially, I'll focus on the latest refactor of the apkg command runner to use asyncio.create_subprocess_exec() and asyncio.create_subprocess_shell() functions which allows previously unthinkable - capturing and printing stdout/stderr at the same timeIndeed, after many years of development, Python can finally perform the equivalent of unix tee command available to mankind since the dawn of time.